To date, a proposal has been made of an apparatus that extracts an event end point of a sport video that is subjected to digest viewing. In this apparatus, cut points on and after the event start point are detected from the sport video whose event start point has been determined, and the type of the cut length between the detected cut points is classified in accordance with the length of the cut length. Also, an appearance pattern of the type of the cut length is stored in an appearance pattern storage unit in association with the event end point. The appearance pattern storage unit is then referred to, and an event end point corresponding to the appearance pattern of the classified type is extracted, and the extracted event end point is output.